Airhop: React Native implementation of bitchat

[areebahmeddd]

Hey @jackjackbits @callebtc

I'm a final year CS undergrad. Had some time to kill and loved the concept/tech behind bitchat, so I decided to port it to React Native at a 24 hour hackathon. That became Airhop. It's going live on Google Play, App Store and Zapstore this month

I spent the last month getting protocol compatibility right rather than just rebuilding the app. It's tested against your implementations, including the v2 packet format, Noise XX/X, GCS gossip sync, geohash presence, courier envelopes, QR-verifying system, private groups, board posts, push-to-talk, mesh bridge, internet gateway, announce signing, key pinning and the peer binding changes from #1432

I had an AI agent monitor your commits, letting me review and pull relevant protocol changes into Airhop as they landed on main. Airhop also addresses a number of open items on your tracker :

• Wifi Aware Support #907
• Feature Request: Confirmation Step or Config Option for Emergency Wipe #152
• Chore: Consolidate base64url helpers into a single utility #642
• Android Parity: Packet Signing and Private File Transfer Fixes #933
• Security UX: Default View Should Not Be the Unencrypted Mesh Channel #1064
• fix: Stop a spoofed fragment from truncating an in-flight reassembly #1515
• Arti: No pluggable transports, unsafe as enabled default #1549
• Feature: Add per-conversation notification controls #1594
• Remove the public-channel screenshot broadcast #1596
• feat: surface private media decode failures in the DM thread #1609
• iOS/Android disagree on max expanded (decompressed) payload size — 1.13 MiB vs 10 MiB, silent drop on iOS #1618
• fix(dm): NFC-normalize nicknames when consolidating geo DM threads #1625
• fix(composer): match @mentions across Unicode NFC/NFD forms #1626
• fix: exclude blocked people from geohash participant counts #1637

Upstream fixes from Airhop:

#1587
#1631
permissionlesstech/bitchat-android#863

The project is also pretty heavily tested. There are around 3,100 unit and simulation tests, along with real world testing between Airhop ↔ Airhop and Airhop ↔ bitchat covering discovery, messaging, voice and file transfers (they're wire-compatible and use the same service UUIDs)

Every release includes source hashes and build provenance as well

The main focus was on:

• UI/UX: I leaned heavily on Laws of UX and Nielsen's usability heuristics because I believe great privacy tools should be easy to use, not something people have to figure out

• Transports: Extended local connectivity with Wi-Fi Aware on Android and iOS and mDNS/TCP LAN discovery, giving peers more ways to find and connect to each other beyond Bluetooth and the existing mesh transports

• Forward secrecy: Double Ratchet on top of Noise for mesh DMs using packet type 0x12 between Airhop peers. Bitchat peers simply ignore it. The root key comes from the Noise XX exporter secret rather than the handshake transcript since handshakes are flooded at TTL 7. This is complementary to Add double-ratchet encryption for Nostr relay DMs #1107 for Nostr relay messaging

• Payments: Dedicated Cashu wallet with offline ecash transfers, Lightning deposits and withdrawals, Nutzaps, DLEQ verification and recovery support

• Tor: Extended the existing Arti integration with bridge support for obfs4 and Snowflake along with bridge fallback and configuration for networks where direct Tor connections are blocked or restricted

• Documentation: I regularly update the readme, known issues and implementation gaps (docs/). I also maintain a landing page with a transparent overview of the project and architecture :

  • https://airhop.1mindlabs.org
  • https://airhop.1mindlabs.org/architecture

Thanks for building bitchat and keeping it public domain. Like bitchat, Airhop is free and MIT licensed. I would love for you to try it and share feedback. I'm happy to answer questions, take corrections or send patches upstream if anything is useful 🙂

Note : This isn't meant to be a cheap, one-shot AI copy of existing work. I plan to build a full ecosystem around it across desktop, smartwatches and CLI while staying in sync with bitchat. If you like the direction of the project, consider supporting it by starring the repo, contributing or donating

[areebahmeddd]

Since this is a third-party implementation of the bitchat wire protocol, we originally used 0x2a/0x2b for extensions but are moving them after seeing the spray-ack reservation in MessageType.swift

The same gap exists in the ANNOUNCE TLV namespace: Bitle already ships 0xB0 and 0xB1 for firmware version and a relay-role marker, and we read the role bit so relay nodes render as infrastructure instead of as a person nobody can message. Reserving a block in both namespaces, TLV as well as message type, would let work like that land without every client guessing which tags upstream will never allocate.

Would you consider reserving a small block, say 0x50–0x5f, for third-party extensions and documenting it as never allocated upstream? This would give other clients room to add protocol features without conflicting with upstream allocations

[alonegggirl628-sketch]

Really impressive work, especially the focus on protocol compatibility and security. The testing depth and real-world Airhop bitchat coverage make this stand out. 1ppp game aside, the forward secrecy and UX improvements are especially interesting. Excited to see how the ecosystem develops from here!

[shibohaneef-stack]

Really impressive work on Airhop and the protocol compatibility. The focus on privacy, testing, and cross-platform support stands out. Excited to see it launch on Google Play and F-Droid. Wishing the project continued success! z777

[suprabhaat]

NOOOOO! flutter is better

[vehsjsbshshshshs-crypto]

Really impressive work on Airhop and the protocol compatibility! The focus on privacy, testing, and usability is especially interesting. Excited to see how the ecosystem develops from here. Also curious about the payment side, including Royal X Casino